Title: Skew distributions in model-based clustering
Authors: Sharon Lee - University of Queensland (Australia) [presenting]
Abstract: The past decade has seen increasing use of flexible distributions that can handle skewness in the data. The literature now offers a wide variety of non-normal and asymmetric distributions with different characterizations and properties. These distributions are motivated by and are suitable for different applications. The aim is to present a survey of some popular skew distributions adopted in the model-based clustering literature, including the class of skew symmetric distributions, the multiple scaled distributions, as well as those obtained by transformation. Their formulations, properties, and advantages will be briefly discussed. Some simulations will be presented to illustrate their abilities in modelling distinct types of skewness in the data.